Background:  An estimated 25% of HIV infections in the United States and 22 to 44% in New York City (NYC) are undiagnosed or unreported, so information on HIV risk behaviors for the general population is difficult to obtain. This study estimates the prevalence of diagnosed and undiagnosed or unreported HIV infection in NYC and compares risk behaviors among NYC’s general adult population of infected and uninfected persons.

Methods:  The population-based NYC Health and Nutrition Examination Survey (HANES) was conducted in 2004 among non-institutionalized adults aged ≥20 years. To identify previously reported HIV infection, we matched NYCHANES with NYC’s HIV/AIDS Registry System (HARS) of all persons with reported HIV infection, using name, date of birth and social security number. Then with consent, 1626 interviewed participants’ archived blood was tested anonymously for HIV. Weighted analyses accounted for complex sampling, nonparticipation and post-stratification to represent the total NYC adult population. Prevalence estimates and 95% confidence intervals were calculated for HIV infection, undiagnosed or unreported HIV, high-risk activities, and perceived risk of having HIV/sexually transmitted disease (STD).

Results:  HIV prevalence among NYC adults was 1.4% (n = 21; 95%CI, 0.6 to 2.2); 5% (95%CI, 0.7 to 29.9) of HIV-infected adults were not reported in HARS, representing undiagnosed or unreported HIV. High-risk activities among HIV-infected adults, including having multiple sex partners (past year) (26%) and inconsistent condom use by people with multiple partners (past month) (33%), did not differ significantly from those of uninfected adults. Needle use history did differ between the infected and uninfected (21 vs 1%, p <0.01). Among people engaging in high-risk behaviors, while 92% reported low/no risk of currently having HIV/STD, uninfected persons were more likely to perceive minimal risk compared with infected persons (p <0.01).

Conclusions:  Although the exclusion of institutionalized groups is a limitation, a smaller proportion of HIV-infected NYC adults was undiagnosed or unreported than previously estimated. Ongoing risky behaviors indicate that more effective prevention interventions are needed, particularly for HIV-infected persons. Low risk perception among adults engaged in high-risk activities highlights the limitations of risk-based testing. Broader, routine screening will enable identification of infected persons not targeted for testing or who do not believe they are at risk.